Food energy values used for nutrition labelling and other purposes are traditionally based on the metabolisable energy (ME) standard, which has recent support from Warwick & Baines (2000). By reference to current practices and published data, the present review critically examines the ME standard and support for it. Theoretical and experimental evidence on the validity of ME and alternatives are considered. ME and alternatives are applied to 1189 foods to assess outcomes. The potential impact of implementing a better standard in food labelling, documentation of energy requirements and food tables, and its impact on users including consumers, trade and professionals, are also examined. Since 1987 twenty-two expert reviews, reports and regulatory documents have fully or partly dropped the ME standard. The principal reason given is that ME only approximates energy supply by nutrients, particularly fermentable carbohydrates. ME has been replaced by net metabolisable energy (NME), which accounts for the efficiency of fuel utilisation in metabolism. Data collated from modern indirect calorimetry studies in human subjects show NME to be valid and applicable to each source of food energy, not just carbohydrates. NME is robust; two independent approaches give almost identical results (human calorimetry and calculation of free energy or net ATP yield) and these approaches are well supported by studies in animals. By contrast, the theoretical basis of ME is totally flawed. ME incompletely represents the energy balance equation, with substantial energy losses in a missing term. In using NME factors an account is made of frequent over-approximations by the ME system, up to 25 % of the NME for individual foods among 1189 foods in British tables, particularly low-energy-density traditional foods. A new simple general factor system is possible based on NME, yet the minimal experimental methodology is no more than that required for ME. By accounting for unavailable carbohydrate the new factor system appears as specific to foods as the USA's food-specific Atwater system, while it is more representative of energy supply from food components. The NME content of foods is readily calculable as the sum from fat (37 kJ/g), protein (13 kJ/g), available carbohydrate (16 kJ/g), fully-fermentable carbohydrate (8 kJ/g), alcohol (26 kJ/g) and other components. Obstacles to the implementation of NME appear to be subjective and minor. In conclusion, the ME standard is at best an approximate surrogate for NME, and inadequately approximates food energy values for the purpose of informing the consumer about the impact on energy balance of the energy supply for equal intake of individual foods. NME is superior to ME for nutrition labelling and other purposes.